Liquid AI发布LFM3-Audio-2B端侧音频模型
New audio model: LFM3-Audio-2B
One more great model from Liquid AI!
Liquid AI 又一款出色的模型！
Good for edge:
适用于边缘计算：
LFM3-Audio-2B is a compact, fully featured audio model that runs locally on laptops and edge devices. It takes text or audio as input and responds with text, speech, or both.
LFM3-Audio-2B 是一款紧凑且功能齐全的音频模型，可在笔记本电脑和边缘设备上本地运行。它接受文本或音频作为输入，并以文本、语音或两者兼有的形式响应。
- Voice chat: ask questions aloud and receive streaming replies in text and speech. For text-only responses, the model can reason before answering.
- Multilingual speech: transcribe, synthesize, and translate speech across 18 languages.
- Audio understanding and reasoning: interpret recordings and answer questions about the speech, sounds, and events they contain.
- Tool calling: turn spoken requests into tool calls to control apps and devices, with text or speech alongside the calls.
- 语音聊天：大声提问，并接收以文本和语音形式流式传输的回复。对于仅文本的响应，模型可以在回答之前进行推理。
- 多语言语音：对 18 种语言的语音进行转录、合成和翻译。
- 音频理解与推理：解读录音内容，并回答关于其中包含的语音、声音和事件的问题。
- 工具调用：将语音请求转化为工具调用，以控制应用程序和设备，并在调用时附带文本或语音。
Capabilities
能力
- Speech recognition (ASR)
- Real-time speech recognition (LiveASR)
- Text-to-speech synthesis (TTS)
- Voice chat with spoken and text responses
- Speech-to-text chat with optional reasoning
- Audio understanding
- Speech translation
- Voice-activated tool calling
- 语音识别 (ASR)
- 实时语音识别 (LiveASR)
- 文本转语音合成 (TTS)
- 支持语音和文本响应的语音聊天
- 语音转文本聊天，可选推理功能
- 音频理解
- 语音翻译
- 语音激活的工具调用
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力